Integrated single-cell analysis of human atherosclerotic lesion

Short description

This report shows the re-analysis workflow of publicly available scRNA-seq data collected from the cells of human atherosclerotic lesion or adjacent artery tissue.

Data sources

Authors	Year	Sample details	Number of samples	Sample groups	GSE ID	Publication
Alsaigh T. et al.	2020	carotid artery	6	Patient-matched calcified atherosclerotic plaques (n=3), and proximal adjacent tissues (n=3)	GSE159677	https://pubmed.ncbi.nlm.nih.gov/36224302/
Pan H. et al.	2020	carotid artery	3	Atherosclerotic carotid arteries from patients undergoing endarterectomy (n=3)	GSE155512	https://pubmed.ncbi.nlm.nih.gov/32962412/
Wirka R.C. et al.	2019	coronary artery	8	Atherosclerotic coronary arteries from explanted hearts of patients undergoing heart transplantation (n=4), with replicates	GSE131778	https://pubmed.ncbi.nlm.nih.gov/31359001/

Data analysis

Analysis using Seurat (version 4.9.9.9091).

Sample (library) information

Clinical characteristics of samples and patients

Technical characteristics of DNA libraries pooled by sample while processing via CellRanger

Initial QC metrics estimation

• Number of genes detected per cell (nFeature_RNA)
• Number of gene counts per cell (nCount_RNA)
• Fraction (percent) of the most expressed gene (percent_largest_gene)
• Fraction (percent) of mitochondrial genes (percent_mito)
• Fraction (percent) of ribosomal protein genes (percent_ribo)
• Fraction (percent) of haemoglobin genes (percent_hb)
• Fraction (percent) of Malat1 - high levels can point to ambient RNA contamination (percent_malat1)
• Estimate of transcriptome complexity in each cell, a ratio of nGenes over nUMIs (novelty_score)

Mito-ribo ratio (MRR) is rather useful estimation of cell quality. Barcodes with higher MRR (more mitochondrial and less ribosomal genes expressed) usually appear to be diying or apoptotic cells, or cell debris.

Here are dashed lines denote approximate thresholds of the relevant parameters picked “by eye”: - nGenes = [200; 4500], - MRR ≤ 0.5.

Another useful characteristics is an estimation of DNA library complexity called a novelty score (NS). Sometimes one can detect contamination with low complexity cell types like red blood cells via this metric. Usually, it is expected to be above 0.80.

What are the top 10 mostly expressed gene in every sample (library)?

A high percentage of MALAT1 and MT genes point to both bad quality (dying, apoptotic) cells or contamination by ambient RNA.

As we see above, MALAT1 is highly presented mostly in the cells with high MRR. But also highly enriched in the cells with MRR <0.5 but with not so many (<2000) genes detected per cell. Hence, the presence of ambient RNA is expected due to possibly abundant presence of debris in atherosclerotic lesions.

Let’s start by correcting for any potential RNA contamination from the surroundings, and afterward, we’ll implement quality control filtering.

Correction for ambient RNA contamination

The presence of cell-free mRNA contamination within the input solution is commonly referred to as “the soup,” which originates from the process of cell lysis. We will check for that and correct using SoupX package.

The top 20 genes with highest expression in background (useful to pick “soup”-specific genes). (These are often enriched for ribosomal proteins).

Automatically estimated ambient RNA total contamination rates

The top 15 genes set to zero in some fraction of cells after SoupX correction.

We observe that certain genes, which have a high level of expression in the ‘soup,’ had their expression levels either reduced to zero or decreased by one or more orders of magnitude.

Conversely, the impact of SoupX correction on certain marker genes for the main cell types is not as dramatic.

Filtering low quality barcodes

There might be different “bad cells” (barcodes): empty drops, fragments of cells, dead cells, etc.

There are several parameters we can explore for cells (barcodes) quality assessment.

• Number of detected genes (NDG, nFeature_RNA): in the original papers, limited between 200 and 4000 genes (Alsaigh, Pan), or between 500 and 3500 (Wirka). Let’s start with 200 detected genes as minimum, and 4500 as maximum.

• Number of gene counts (Number of UMIs) (NUMI, nCount_RNA): usually limited between 500 and 50000 counts. This cutoff mentioned only in Pan et al. paper, number of UMIs is up to 20’000. We will use only lower cutoff of 500 UMIs and set the upper threshold later not to overestimate a doublet rate further.

• Mitochondrial gene counts ratio (MTP, percent_mito): usually must be less than 10%. Ppers by Pan et al and Alsaigh et al use a 10% cutoff. In Wirka et al study it is more stringent (< 7.5%). Let’s first use more relaxed 10% since we define an additional cutoff of MRR <0.5.

• Hemoglobin gene counts ratio (HBP, percent_hb): must be less than 1% which helps to dispose of red blood cell contamination. There is almost no cells in analysed libraries with noticeable levels of HB genes.

• Complexity: novelty score (NS, novelty_score): recommended to be more than 0.8.

How many cells pass these criteria?

Let’s take a look at QC metrics after SoupX correction and estimate how many cells pass defined criteria.

sample_id	N_Cells	ALL	Passed NDG cutoff	Passed NUMI cutoff	Passed NS cutoff	Passed MTP cutoff	Passed MRR cutoff	Passed HBP cutoff	Passed all cutoffs
A.GSM4837523	11047	10081	10971 (99.3%)	11034 (99.9%)	10962 (99.2%)	10350 (93.7%)	10232 (92.6%)	11046 (100%)	10081 (91.3%)
A.GSM4837524	3765	3069	3667 (97.4%)	3760 (99.9%)	3685 (97.9%)	3164 (84%)	3111 (82.6%)	3765 (100%)	3069 (81.5%)
A.GSM4837525	16040	13613	15961 (99.5%)	16021 (99.9%)	15522 (96.8%)	14383 (89.7%)	14410 (89.8%)	16040 (100%)	13613 (84.9%)
A.GSM4837526	5590	4725	5558 (99.4%)	5588 (100%)	5432 (97.2%)	5021 (89.8%)	4926 (88.1%)	5590 (100%)	4725 (84.5%)
A.GSM4837527	12531	10515	12335 (98.4%)	12522 (99.9%)	12468 (99.5%)	10944 (87.3%)	10887 (86.9%)	12530 (100%)	10515 (83.9%)
A.GSM4837528	3404	2757	3352 (98.5%)	3399 (99.9%)	3373 (99.1%)	2921 (85.8%)	2801 (82.3%)	3404 (100%)	2757 (81%)
P.GSM4705589	3441	2702	3144 (91.4%)	3437 (99.9%)	3286 (95.5%)	2954 (85.8%)	2925 (85%)	3441 (100%)	2702 (78.5%)
P.GSM4705590	4692	3617	4319 (92.1%)	4689 (99.9%)	4407 (93.9%)	3934 (83.8%)	4025 (85.8%)	4692 (100%)	3617 (77.1%)
P.GSM4705591	3138	2768	3103 (98.9%)	3134 (99.9%)	3127 (99.6%)	2873 (91.6%)	2809 (89.5%)	3138 (100%)	2768 (88.2%)
W.GSM3819856	1828	1528	1825 (99.8%)	1822 (99.7%)	1580 (86.4%)	1792 (98%)	1778 (97.3%)	1828 (100%)	1528 (83.6%)
W.GSM3819857	731	594	731 (100%)	727 (99.5%)	624 (85.4%)	716 (97.9%)	706 (96.6%)	731 (100%)	594 (81.3%)
W.GSM3819858	2225	2007	2221 (99.8%)	2210 (99.3%)	2225 (100%)	2049 (92.1%)	2044 (91.9%)	2225 (100%)	2007 (90.2%)
W.GSM3819859	1975	1789	1974 (99.9%)	1959 (99.2%)	1975 (100%)	1838 (93.1%)	1829 (92.6%)	1975 (100%)	1789 (90.6%)
W.GSM3819860	3012	2872	3008 (99.9%)	3005 (99.8%)	2938 (97.5%)	2959 (98.2%)	2963 (98.4%)	3012 (100%)	2872 (95.4%)
W.GSM3819861	3192	3026	3180 (99.6%)	3178 (99.6%)	3130 (98.1%)	3108 (97.4%)	3124 (97.9%)	3192 (100%)	3026 (94.8%)
W.GSM3819862	2907	2754	2904 (99.9%)	2896 (99.6%)	2846 (97.9%)	2836 (97.6%)	2839 (97.7%)	2906 (100%)	2754 (94.7%)
W.GSM3819863	2547	2408	2541 (99.8%)	2543 (99.8%)	2518 (98.9%)	2472 (97.1%)	2450 (96.2%)	2546 (100%)	2408 (94.5%)

Doublet/multiplet detection

Doublets/multiplets are defined as two or more cells that are sequenced under the same cellular barcodec. They can be formed from the same (homotypic) or different (heterotypic) cell types. Their identification is crucial as they are most likely misclassified and can lead to distorted downstream analysis steps.

To detect putative doublets, will use three R packages: - DoubletFinder - scDblFinder. - Scrublet, adopted for run from R here

Analysed dataets are obtained after scRNA-seq of cell suspension pre-sorted by cell size and viability. They must have contain rather low percent of droplets with doublet/multiplet cells. Doublet rate about 3% is a maximum remarked in related publications (in Alsaigh et al paper). We set this rate higher, up to 8%, to certainly filter out putative doublets.

Let’s do initial merging of sample datasets to visually compare the efficiency of doublet detection methods and QC.

QC metrics of initially merged dataset:

Consensus between different doublet detection methods - how many cells considered as doublets are overlapped?

                     DoubletFinder Doublet Singlet
Scrublet scDblFinder                              
Doublet  Doublet                       395     779
         Singlet                        13     102
Singlet  Doublet                       487    3042
         Singlet                      4064   61869

Since we have enough cells, let’s discard all barcodes considered as doublet by any method.

Additionally, we employ a filtering criterion based on the number of reads per gene (nCount_RNA should be less than 15000) to mitigate the potential presence of putative doublets in the dataset. Additionally, we will filter out all CD45+ cells located within non-leikosyte superclusters (endothelial and smooth muscle cells).

Number of cells with gene counts outside the thresholds (nCount_RNA > th_nch) in total and by library (sample):

Total number of cells with gene counts > 15000 

FALSE  TRUE 
67336  3415 


Number of cells with gene counts > 15000 by sample 
              
               FALSE  TRUE
  A.GSM4837523  9208   871
  A.GSM4837524  2957   111
  A.GSM4837525 13173   432
  A.GSM4837526  4539   185
  A.GSM4837527  9785   725
  A.GSM4837528  2575   179
  P.GSM4705589  2416   285
  P.GSM4705590  3211   405
  P.GSM4705591  2655   109
  W.GSM3819856  1518     4
  W.GSM3819857   591     0
  W.GSM3819858  1997     3
  W.GSM3819859  1774     5
  W.GSM3819860  2844    23
  W.GSM3819861  3003    14
  W.GSM3819862  2738    10
  W.GSM3819863  2352    54

To have more cues about cell types in detected clusters in initially merged dataset, we will use a pre-defined list of markers for anticipated cell types:

We suggest the following main cell types

Cell type	Cluster
CD45+ cells (leukocytes)	0,1,2,6,8,9,14,19,20,21,24
Endothelial cells	4,10,11,23
Smooth muscle cells + Fibroblasts	3,5,7,12,13,15,16,17,18,22

Detected CD45+ cells in non-leukocytes clusters:

Number of CD45+ cells in endothelial clusters (EC): 
CD45-EC CD45+EC 
   9271     392 

Number of CD45+ cells in smooth muscle clusters (SMC): 
CD45-SMC CD45+SMC 
   20835      739 

The number of detected genes in these cells

The number of detected genes in these suspicious cell groups is higher which may indicate the higher probability these “cells” are doublets that are not detected by either tool used for this purpose.

Thus, if assuming that transition of VSMC and EC to CD45+ myeloid cell is not possible (but we cannot say the same about the reverse transition), we may estimate the observed rate of heterotypic leukocyte-derived doublets is about 3-4%.

Let’s estimate the number of cells passing all filters.

Total number of cells passed after doublet detection and 'gene counts' cutoff:
Discarded    Passed 
    11255     59496 

Number of cells in every samples passed after doublet detection and 'gene counts' cutoff:              
               Discarded Passed
  A.GSM4837523      2008   8071
  A.GSM4837524       372   2696
  A.GSM4837525      2302  11303
  A.GSM4837526       625   4099
  A.GSM4837527      1986   8524
  A.GSM4837528       443   2311
  P.GSM4705589       502   2199
  P.GSM4705590       733   2883
  P.GSM4705591       386   2378
  W.GSM3819856       165   1357
  W.GSM3819857        68    523
  W.GSM3819858       275   1725
  W.GSM3819859       205   1574
  W.GSM3819860       309   2558
  W.GSM3819861       310   2707
  W.GSM3819862       283   2465
  W.GSM3819863       283   2123

Since the priority of this study is plaque cells, we will also remove the samples of proximal adjacent tissue (PA samples: GSM4837524, GSM4837526, GSM4837528) from Alsaigh et al. dataset which seem to have some fraction of adventitial cells.

Data normalisation and clustering

Let’s proceed to normalize the remaining samples utilizing standard LogNorm workflow in Seurat and then integrate them using Harmony.

Cell clusters:

QC Metrics

Cell clusters dendrogram:

Cell clusters by source:

Cluster annotation

First look how are expressed some pre-selected marker genes specific for atherosclerotic arteries

Top markers (up-regulated genes) for cell clusters

The expression of cluster-specific marker genes (top 10 for each cluster) identified by MAST method.

Top 5 marker genes:

Automatic cell type annotation using

• SingleR R package and human RNA-seq reference via celldex package.

	C0	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20	C21	C22	C23
Astrocyte	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	24	1	0	0	0
B_cell	0	0	0	0	0	0	1	0	2271	0	0	0	0	0	1	0	0	1	6	0	1	1	0	18
Chondrocytes	0	0	2104	3	0	1415	4	0	2	0	0	156	974	136	0	472	0	0	28	22	0	93	2	0
CMP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	627	0	0	0	0	0	0	0	0	0
DC	0	0	1	4	177	1	0	132	1	0	7	1	0	0	3	0	33	66	0	0	2	0	0	4
Endothelial_cells	0	0	8	4292	0	0	0	0	0	0	0	7	0	0	2	0	0	0	0	1	3	2	257	0
Epithelial_cells	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Fibroblasts	0	0	96	3	0	208	0	0	0	0	0	54	28	118	1	36	0	0	26	5	1	21	0	0
GMP	0	0	0	0	0	0	0	0	1	0	0	0	0	0	41	0	0	0	1	0	4	0	0	12
Hepatocytes	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	0
HSC_-G-CSF	0	0	0	0	0	0	1	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0
HSC_CD34+	0	0	0	3	0	0	0	1	2	0	0	0	0	0	9	0	0	0	0	0	0	0	2	0
Macrophage	0	0	0	1	2654	2	1	1450	0	0	35	0	0	0	16	0	376	24	3	0	55	0	1	0
Monocyte	0	0	0	0	1385	0	0	919	3	0	1937	0	0	0	69	0	336	391	1	0	33	0	0	34
MSC	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0
Neurons	0	0	0	1	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	288	1	1	0	0
Neutrophils	0	0	0	0	0	0	0	2	1	0	34	0	0	0	5	0	0	0	2	0	0	0	0	0
NK_cell	324	18	0	0	0	0	23	3	6	1537	0	0	0	0	96	0	0	0	15	0	33	0	0	1
Osteoblasts	0	0	18	0	0	15	0	0	0	0	0	3	16	0	0	1	0	0	0	0	0	5	0	0
Pre-B_cell_CD34-	4	3	0	1	0	0	10	5	9	1	0	0	0	0	13	0	0	0	3	0	0	0	0	42
Pro-B_cell_CD34+	0	0	0	0	0	0	0	0	7	0	0	0	0	0	1	0	0	0	0	0	3	0	0	0
Pro-Myelocyte	0	0	0	0	0	0	0	0	0	0	0	0	0	0	4	0	0	0	0	0	0	0	0	0
Smooth_muscle_cells	0	0	791	13	0	1677	0	0	0	0	0	651	628	624	0	222	0	0	42	35	0	133	5	0
T_cells	5845	4841	0	1	0	1	2930	5	6	562	1	0	0	0	6	0	0	0	282	0	280	2	1	0
Tissue_stem_cells	0	0	1652	20	0	822	6	0	2	0	0	1080	303	238	1	110	0	0	71	78	0	90	5	0

	C0	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	C18	C19	C20	C21	C22	C23
B cell	90	90	0	1	0	1	242	5	1420	6	0	0	0	0	2	0	0	4	41	0	6	0	0	63
endothelial cell	0	0	0	3835	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	3	0	146	0
endothelial cell of artery	0	0	0	224	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	10	0
fibroblast	0	0	1	4	0	1470	1	0	0	0	0	6	117	1051	0	37	0	0	112	32	0	7	1	0
macrophage	81	24	3	62	4215	4	156	2506	579	9	2013	11	0	0	100	0	745	478	84	9	107	29	38	18
mast cell	0	0	0	0	0	0	0	0	0	0	0	0	0	0	790	0	0	0	0	0	0	0	0	0
mature NK T cell	103	6	0	0	0	0	8	0	1	1406	0	0	0	0	0	0	0	0	1	0	25	0	0	0
pericyte	0	0	21	3	0	9	0	0	0	0	0	900	1	0	0	1	0	0	0	16	0	0	0	0
plasma cell	0	0	1	0	1	0	0	4	309	0	1	1	0	0	1	0	0	0	3	0	2	2	1	30
smooth muscle cell	1	0	4644	213	0	2656	12	0	1	0	0	1035	1831	65	2	803	0	0	44	397	3	308	76	0
T cell	5898	4742	0	1	0	1	2558	2	1	679	1	0	0	0	0	0	0	0	195	0	271	2	1	0

Main immune cell type gene signatures from Zernecke A. et al, 2023

Some gene markers of lipid uptake and foam cells

Clusters C7 (macrophages) and C11 (SMCs) have APOE, CD36, FABP4, FABP5 highly co-expressed which are related to lipid uptake and storage, and may point to a possible “fomay” state of these cells.

Components of complement system

Some markers of osteoblasts

Cell numbers in manually annotated cell types and subtypes

	T cell	Smooth muscle cell	Endothelial cell	Macrophage	B cell	Monocyte	Fibroblast	Mast cell	Dendritic cell	Fibroblast/Mixed	Neuron	Proliferating immune cells	Plasma cells
A.GSM4837523	2839	1023	273	2726	35	804	1	64	108	73	0	110	15
A.GSM4837525	6263	1430	1182	695	1047	218	49	135	38	148	0	80	18
A.GSM4837527	3990	2419	76	796	129	509	15	231	95	89	0	122	53
P.GSM4705589	26	1908	68	103	0	13	6	61	3	2	0	8	1
P.GSM4705590	192	995	588	713	26	137	7	115	82	2	0	17	9
P.GSM4705591	392	903	265	578	12	76	15	46	68	6	1	11	5
W.GSM3819856	462	200	87	202	297	19	38	6	10	12	10	14	0
W.GSM3819857	95	122	127	55	73	9	25	3	4	7	2	0	1
W.GSM3819858	296	787	178	30	7	21	192	3	1	35	174	1	0
W.GSM3819859	285	743	61	81	9	31	174	1	2	22	162	3	0
W.GSM3819860	396	755	529	343	198	42	186	21	13	24	36	15	0
W.GSM3819861	460	786	584	288	261	51	170	14	13	26	42	12	0
W.GSM3819862	397	705	518	328	215	42	177	13	14	21	25	10	0
W.GSM3819863	19	1126	80	540	2	43	61	182	31	13	3	14	9

	T cell (Cytotoxic)	T cell (Activated)	SMC (Contractile)	Endothelial cell (C4)	Macrophage (Inflammatory)	SMC (Fibromyocyte C5)	T cell (Memory)	Macrophage (Foamy)	B cell	T cell (T/NK cell)	Monocyte	SMC (Foamy)	SMC (Fibromyocyte C12)	Fibroblast	Mast cell	SMC (Osteochondrogenic)	Macrophage (C16)	Dendritic cell	Fibroblast/Mixed (C18)	Neuron	Proliferating immune cell	SMC (Undefined C21)	Endothelial cell (C22)	Plasma cell
A.GSM4837523	941	903	386	195	401	214	438	1605	35	557	804	28	78	1	64	299	720	108	73	0	110	18	78	15
A.GSM4837525	2750	1834	208	1177	551	573	1131	138	1047	548	218	361	176	49	135	68	6	38	148	0	80	44	5	18
A.GSM4837527	1283	1196	1115	59	614	731	1047	172	129	464	509	113	323	15	231	26	10	95	89	0	122	111	17	53
P.GSM4705589	11	12	961	39	92	489	2	11	0	1	13	116	304	6	61	24	0	3	2	0	8	14	29	1
P.GSM4705590	106	53	288	509	609	250	18	104	26	15	137	120	124	7	115	197	0	82	2	0	17	16	79	9
P.GSM4705591	157	129	274	238	485	218	54	89	12	52	76	83	139	15	46	169	4	68	6	1	11	20	27	5
W.GSM3819856	160	140	43	87	151	58	47	51	297	115	19	56	31	38	6	2	0	10	12	10	14	10	0	0
W.GSM3819857	32	28	25	120	45	50	5	10	73	30	9	31	11	25	3	0	0	4	7	2	0	5	7	1
W.GSM3819858	84	106	199	177	18	307	39	12	7	67	21	148	103	192	3	11	0	1	35	174	1	19	1	0
W.GSM3819859	92	96	184	58	67	295	33	14	9	64	31	161	84	174	1	8	0	2	22	162	3	11	3	0
W.GSM3819860	179	108	142	528	278	274	53	63	198	56	42	195	120	186	21	6	2	13	24	36	15	18	1	0
W.GSM3819861	189	144	185	582	227	256	55	60	261	72	51	222	98	170	14	8	1	13	26	42	12	17	2	0
W.GSM3819862	182	108	139	514	265	226	50	62	215	57	42	219	104	177	13	8	1	14	21	25	10	9	4	0
W.GSM3819863	7	5	521	60	413	200	5	126	2	2	43	100	254	61	182	15	1	31	13	3	14	36	20	9

THOR target genes

How are THOR genes expressed in cell clusters?

References

This workflow is inspired by several sources:

• “Best practices for single-cell analysis across modalities” paper by Heumos et al (2023)

• “Single-cell RNA-seq: Quality Control Analysis”, teaching materials at the Harvard Chan Bioinformatics Core

• Quality Control chapter in “Single cell best practices” handbook

• Single-cell data analysis pipeline elaborated by CellGenIT group

R packages used

sessionInfo()

R version 4.2.3 (2023-03-15)
Platform: x86_64-conda-linux-gnu (64-bit)
Running under: Ubuntu 22.04.1 LTS

Matrix products: default
BLAS/LAPACK: /home/amarkov/miniconda3/envs/r4.2/lib/libopenblasp-r0.3.24.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats4    stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] harmony_1.0.1               Rcpp_1.0.9                  UCell_2.2.0                 DT_0.27                    
 [5] khroma_1.9.0                MetBrewer_0.2.0             RColorBrewer_1.1-3          gdata_2.19.0               
 [9] Matrix_1.6-1.1              janitor_2.2.0               lubridate_1.9.3             forcats_1.0.0              
[13] stringr_1.5.0               dplyr_1.1.3                 purrr_1.0.2                 readr_2.1.4                
[17] tidyr_1.3.0                 tibble_3.2.1                ggplot2_3.4.3               tidyverse_2.0.0            
[21] patchwork_1.1.3             gridExtra_2.3               scrubletR_0.1.0             scDblFinder_1.10.0         
[25] DoubletFinder_2.0.3         SoupX_1.6.2                 scCustomize_1.1.1           Seurat_4.9.9.9067          
[29] SeuratObject_4.9.9.9091     sp_2.1-0                    SingleCellExperiment_1.20.0 SummarizedExperiment_1.28.0
[33] Biobase_2.58.0              GenomicRanges_1.50.0        GenomeInfoDb_1.34.9         IRanges_2.32.0             
[37] S4Vectors_0.36.0            BiocGenerics_0.44.0         MatrixGenerics_1.10.0       matrixStats_1.0.0          

loaded via a namespace (and not attached):
  [1] ggprism_1.0.4             rtracklayer_1.58.0        scattermore_1.2           ragg_1.2.5               
  [5] bit64_4.0.5               knitr_1.44                irlba_2.3.5.1             DelayedArray_0.24.0      
  [9] data.table_1.14.8         RCurl_1.98-1.12           doParallel_1.0.17         generics_0.1.3           
 [13] ScaledMatrix_1.6.0        RhpcBLASctl_0.23-42       cowplot_1.1.1             RANN_2.6.1               
 [17] future_1.33.0             bit_4.0.5                 tzdb_0.4.0                spatstat.data_3.0-1      
 [21] httpuv_1.6.11             viridis_0.6.4             xfun_0.40                 jquerylib_0.1.4          
 [25] hms_1.1.3                 evaluate_0.22             promises_1.2.1            progress_1.2.2           
 [29] fansi_1.0.5               restfulr_0.0.15           readxl_1.4.3              DBI_1.1.3                
 [33] igraph_1.4.2              htmlwidgets_1.6.2         spatstat.geom_3.2-5       paletteer_1.5.0          
 [37] ellipsis_0.3.2            crosstalk_1.2.0           RSpectra_0.16-1           prismatic_1.1.1          
 [41] deldir_1.0-9              sparseMatrixStats_1.10.0  vctrs_0.6.3               ROCR_1.0-11              
 [45] abind_1.4-5               cachem_1.0.8              withr_2.5.1               progressr_0.14.0         
 [49] vroom_1.6.4               presto_1.0.0              sctransform_0.4.0         GenomicAlignments_1.34.0 
 [53] prettyunits_1.2.0         scran_1.24.1              goftest_1.2-3             cluster_2.1.4            
 [57] ape_5.7-1                 dotCall64_1.0-2           lazyeval_0.2.2            crayon_1.5.2             
 [61] spatstat.explore_3.2-3    labeling_0.4.3            edgeR_3.38.4              pkgconfig_2.0.3          
 [65] nlme_3.1-163              vipor_0.4.5               rlang_1.1.1               globals_0.16.2           
 [69] lifecycle_1.0.3           miniUI_0.1.1.1            fastDummies_1.6.3         rsvd_1.0.5               
 [73] cellranger_1.1.0          ggrastr_1.0.1             polyclip_1.10-6           RcppHNSW_0.5.0           
 [77] lmtest_0.9-40             zoo_1.8-12                beeswarm_0.4.0            ggridges_0.5.4           
 [81] GlobalOptions_0.1.2       png_0.1-8                 viridisLite_0.4.2         rjson_0.2.21             
 [85] bitops_1.0-7              KernSmooth_2.23-22        spam_2.9-1                Biostrings_2.66.0        
 [89] DelayedMatrixStats_1.20.0 shape_1.4.6               parallelly_1.36.0         spatstat.random_3.1-6    
 [93] beachmat_2.14.0           scales_1.2.1              magrittr_2.0.3            plyr_1.8.9               
 [97] ica_1.0-3                 zlibbioc_1.44.0           compiler_4.2.3            dqrng_0.3.1              
[101] BiocIO_1.8.0              clue_0.3-64               fitdistrplus_1.1-11       Rsamtools_2.14.0         
[105] snakecase_0.11.0          cli_3.6.1                 XVector_0.38.0            listenv_0.9.0            
[109] pbapply_1.7-2             MASS_7.3-60               tidyselect_1.2.0          MAST_1.22.0              
[113] stringi_1.7.12            BPCells_0.1.0             textshaping_0.3.6         yaml_2.3.7               
[117] BiocSingular_1.14.0       locfit_1.5-9.7            ggrepel_0.9.3             grid_4.2.3               
[121] sass_0.4.7                tools_4.2.3               timechange_0.2.0          future.apply_1.11.0      
[125] parallel_4.2.3            circlize_0.4.15           rstudioapi_0.15.0         bluster_1.6.0            
[129] foreach_1.5.2             metapod_1.4.0             farver_2.1.1              Rtsne_0.16               
[133] digest_0.6.33             shiny_1.7.5               scuttle_1.6.3             later_1.3.1              
[137] writexl_1.4.2             RcppAnnoy_0.0.21          httr_1.4.7                ComplexHeatmap_2.12.1    
[141] colorspace_2.1-0          XML_3.99-0.14             tensor_1.5                reticulate_1.32.0        
[145] splines_4.2.3             uwot_0.1.16               statmod_1.5.0             rematch2_2.1.2           
[149] spatstat.utils_3.0-3      scater_1.24.0             xgboost_1.7.5.1           systemfonts_1.0.5        
[153] plotly_4.10.2             xtable_1.8-4              jsonlite_1.8.7            R6_2.5.1                 
[157] pillar_1.9.0              htmltools_0.5.6.1         mime_0.12                 glue_1.6.2               
[161] fastmap_1.1.1             BiocParallel_1.32.5       BiocNeighbors_1.16.0      codetools_0.2-19         
[165] utf8_1.2.3                bslib_0.5.1               lattice_0.21-9            spatstat.sparse_3.0-2    
[169] ggbeeswarm_0.7.1          leiden_0.4.3              gtools_3.9.4              survival_3.5-7           
[173] limma_3.54.0              rmarkdown_2.25            munsell_0.5.0             GetoptLong_1.0.5         
[177] GenomeInfoDbData_1.2.9    iterators_1.0.14          reshape2_1.4.4            gtable_0.3.4